Ensembles

Contains implementation of Random Forest, Gradient Boosting algorithms. Also contains a local web server to use these algorithms.

Installation and starting

1. Clone repository
2. Put Dockerfile and files from scripts into the server folder
3. Build docker image:

   sudo bash ./build.sh

4. Run server:

   sudo bash ./run.sh

Usage

1. Type "0.0.0.0:5000" in the search bar in your browser.
2. Choose model to train.

3. Fill parameters, also you can set default parameters by clicking Fill default parameters.

4. Upload csv train and (optional) validation datasets. Each dataset must have at least 1 numeric column and 1 column with target.
5. Specify the target column name.

6. Train model
7. After training you will see results. You can train your model again or you can create a submission on a new (or old) csv file with the same columns (except target).

8. Upload csv file and download submission.csv as a prediction

Link to dockerhub repository: https://hub.docker.com/repository/docker/makriot/ensembles_web_server